SM320F2808-EP: Under 95C storage test the DSP got damaged

Madhu,

For the F2808 the only restriction we have is that no voltage larger than a diode drop(0.7V) be applied to any pin prior to the device being powered, and likewise stay within the recommended operating conditions of the device with respect to VDDIO/VDD voltage.

What was the 3rd part you mention above, is there any possibility the power sequencing of the 3 devices may cause some violation on the F2808 device?  I can't explain how all 3 devices would be damaged in this scenario.  Can you describe the power domains of the 3 devices and how they are related, etc?

Best,

Matthew

Hello Matthew,

Many thanks for your reply. 

DSP is taking input from two ADC128S102. They share a common digital bus. But the chip select is different. What we are noticing is only analog power rail of the DSP had this issue.

Two parts got damaged. One ADC and DSP. If voltage application or power sequencing is an issue, why would the parts get damaged only at high temperature? 

About the power domains, TPS70351 is generating the power to the DSP. 3.3V generated from this device is further used to generate analog 3.3V for the ADC. So the ADCs get digital 3.3V from TPS7031 and analog 3.3V. But there is a slight delay in analog 3.3. Please see the waveform below:

Channel 1 is digital power supply and channel 2 is analog power supply at the ADC. Math channel shows a delay of 237nS between the two.

Do you have the restriction of that no voltage larger than a diode drop(0.7V) be applied to any pin prior to the power to any other analog ICs as well?

Madhu,

Since both devices are getting sourced from same TPS, I don't think there is chance of violation of my above restrictions. 

And, as you said, this is only failing after temp storage I'm assuming it doesn't fail at just room temp or lower storage temp?

When you say that analog rail of DSP is damaged, do you see higher current than normal, or is there a analog(pin/pins) that no longer work?

Given we are in both the operational range and storage range of the device, I would not expect the temperature to be a factor here

Analog supply pin of the DSP was shorted to ground. We confirmed by removing the device..

Are there any specific PCB guidelines for meeting high temperature storage requirements for this package?

Madhu,

If you are seeing analog pin of DSP shorted to ground it is usually due to an overvoltage event on one of the analog input pins.  There are ESD protection diodes on each pin(digital and analog) that will begin to turn on ~0.3V and fully conduct at 0.7V above VDDA supply voltage.  If there is enough current drive(I think ~2mA) it will blow the ESD circuit and then we will see the short from power to ground.

From the scope shot, even if there is some delay I think the devices are within 100s of mV of one another, so again, I can't see how the external ADC would apply an out of spec voltage on power up.

Was the board handled between the temp cycling/power in anyway there may have been an ESD event?  How was the heat being generated, was this from forced air thermal stream, or bake over, etc?

Best,

Matthew

Matthew,

Thank you so much. I do believe that it is a diode conduction issue. If we add resistance on each line, will it help?

Madhu,

Yes, if you can limit the current injection to the device during the over/under voltage to the level I mention previously it will not damage the device with the ESD conducts.

Best,

Matthew

Thank you again. One last related question. I am trying to reproduce the IC failure. I applied 3V on one of the pins without giving VCC. I am seeing current of about 100mA. But the part did not break. Is there any internal limiting?

Madhu,

While the device may still be functioning, likely there has been damage that will shorten its lifetime, and it will eventually fail.  20mA is the value we give to ensure the integrity of the internal metal system for the life of the device.  The longer you apply this over-voltage, the quicker the device will eventually fail.

Best,
Mathhew